J . Am. Chem. SOC.1990, 112, 8260-8265
8260
roughly the average of the atomic gauge factors. The greater flexibility of the CIAO wave function results in faster convergence. As the CIAO method can thus employ smaller basis sets, it may be ultimately the more efficient method, in spite of the savings offered by the localized techniques. Another conclusion emerging from our calculations is that, unless large basis sets are used, the accuracy of individual tensor components is often lower with the localized methods than with the CIAO. This is true even in the cases where the isotropic average is correctly predicted by IGLO and LORG. By contrast, the tensor components are usually fairly well predicted by the CIAO method. Finally, we would like to emphasize that our implementation of the GIAO method can be routinely used for N M R chemical shift calculations on fairly large molecules, say in the Cl0-C,, range, on departmental workstations and minicomputers. Indeed, all results shown in this paper have been obtained on minicomputers. Our program currently runs on the IBM 4381 under CMS and VSI, on the IBM RT PC workstation under AIX, and on the Celerity 1200 and the Apollo DN10000 workstations under Unix. These versions, including integral, closed-shell SCF, and
chemical shift calculations, are available for a nominal fee from the University of Arkansas. The whole TEXAS system, including many other features, some unique, such as a very compact integrals storage, UNO-CAS gradients, and a very efficient geometry optimization using automatically generated internal coordinates, will also be made generally available in the fall of 1990 from the University of Arkansas. Acknowledgment. We gratefully acknowledge partial support for this project by the National Science Foundation under Grant CHE-8500487. We thank IBM Corp. for financial support and for the loan of an PC-RT workstation. The Apollo DNlOOOO used for the large calculations was purchased with funds provided by N S F under Grant CHE-8814143. We thank Prof. P. R. v. Schleyer for the ab initio geometry of the adamantyl cation, Prof. G. Fogarasi for the ab initio optimized geometry of stilbene, Profs. N . Allison and W. L. Meyer for the MMX optimized geometry of octahydroparacyclophane, and Mr. P. Taylor for the ab initio geometries of oxocane. We are also grateful to one of our reviewers for careful and constructive criticism.
Structure and Energetics of C2H4Br+: Ethylenebromonium Ion vs Bromoethyl Cations Tracy P. Hamilton and Henry F. Schaefer, III* Contribution No. 91 from the Center for Computational Quantum Chemistry, University of Georgia, Athens, Georgia 30602. Received March 8, 1990
Abstract: The cyclic and acyclic isomers of C2H4Br+are compared by using ab initio quantum mechanical techniques, including the use of electron correlation and polarization functions. The cyclic bromonium ion is found to be more stable than the acyclic 1 -bromoethyl cation by 1.5 kcal/mol, in very good agreement with experiment. A transition state for the interconversion of
thcse two forms is reported, the energy barrier being -25 kcal/mol. The relative energies of the cyclic and acyclic minima are remarkably insensitive to basis set and electron correlation effects, validating the results of previous low-level studies. The 2-bromoethyl cation does not exist as a minimum on the potential energy surface and spontaneously collapses to the bromonium ion upon torsion.
Introduction
The addition of Br2 to olefins is usually highly stereospecific, and the postulation in 1937 of a cyclic bromonium ion as an intermediate by Roberts and Kimball’ was a radical suggestion. The detection of these intermediates was to come 30 years later in the pioneering NMR studies in superacid media from Olah’s group.* In the interim, the mechanism for addition of Br2 to double bonds via a bridged intermediate was widely accepted because it neatly accounted for why addition was anti: Br+ is added to the double bond, forming a 3-membered ring, and attack by the remaining Br- must come from the other side. The positive charge in the ethylenebromonium ion does not reside on the Br atom at all but primarily on the carbon atoms. Thus there is no attraction between the Br atom in the ring and Br-, and steric factors are free to exert their full effect. A recent review by Ruasse) gives an account of the bromination of olefins, with consideration of whether bromonium ions or @-bromocarbocations are intermediates for various alkenes. ( I ) Roberts, 1.; Kimball, G.E.J . Am. Chem. SOC.1937, 59, 947. (2) (a) Olah. G.A.; Bollinger, J. M. J . Am. Chem. Soc. 1968.90.947. (b) Olah. G. A.: Bollinger, J. M.; Brinich, J. M. J. Am. Chem. SOC.1968. 90, 2587.
(3) Ruasse. M. F. Acc. Chem. Res. 1990, 23. 87.
The open 2-bromoethyl cation is expected to be unfavorable because of the positive charge residing on a primary carbon: methyl substitution should stabilize this carbonium center relative to the bridged isomer. The positive charge on the carbon in the 1-bromoethyl cation is stabilized by Br, which is able to backdonate electron density. The global energy minimum should be either the open 1-halo form or the onium ion. The trend among halogens from experiment favors the open I-halo isomer for F and CI and bridging for Br and I.4,s There may not even be a cyclic ethylenefluoronium ion. Ion cyclotron resonance experiments form Beauchamp’s group4 found two non-interconverting isomers of C2H4Br+and attributed the spectra to the 3-membered ring being 1.4 kcal/mol lower than the I-bromoethyl cation which was made from a different source. This procedure has been criticized because the two different modes of product production may lead to differing levels of internal energy in the two isomers6 Because of its stereoselectivity, it has recently been reported that C2H4Br+ has been used as a (4) Berman, D. W.; Anicich, V.;Beauchamp, J. L. J . Am. Chem. SOC. 1979, 101, 1239. ( 5 ) Olah, G . A. Halonium Ions; Wiley: New York, 1975. (6) For example, see: Angelini, G.; Speranza, M. J . Am. Chem. Soc. 1981, 103, 3792 and ref 14 therein.
0002-7863/90/ I5 12-8260$02.50/0 0 1990 American Chemical Society
Structure and Energetics of C2H4Br+
R V n
rl
a
co
b Figure I . (a) The donor-acceptor interaction between the filled C2H4 7 orbital and the empty adduct orbital. (b) The interaction between the vacant T * orbital and a filled adduct orbital.
chemical ionization reagent to distinguish geometrical isomers in mass spectrometry studies.' The X-ray structure of a substituted ethylenebromonium ion with a Br,- counterion (formed from bromination of adamantylideneadamantane) has also been determined.s The back-side attack by Br- is sterically impossible, so the bromination stops at the intermediate. The close proximity of the counterion introduces an asymmetry in the bromonium ion: the difference in C-Br distances is 0.078 A but other effects are smaller. In the comparison of the symmetric theoretical geometry with the corresponding experimental one, average values will be used. One issue that arises in the discussion of 3-membered rings is whether the ring is a a-complex or a ?r-complex. Figure 1 shows the two types of bonding. The relationship between r-complexes and true 3-membered rings has been discussed by Dewar and Ford.9 Cremer and Kraka'O also studies this relationship using an analysis of computed electron densities. This complex issue will be addressed later in this paper, keeping in mind that there is no sharp boundary between the two. The only previous theoretical studies of the CzH4Br+system were performed at the S C F level with small basis sets and incomplete optimization." Also, the characterization of stationary points was not feasible at the time of the study, nor was the search for a C I transition state. Given the previous low levels of theory that were used to treat the bromonium ion, the question of Beauchamp's values, and the intervening X-ray structure analysis, we decided to clarify the energetics and structures of the various CzH4Br+isomers. The justification for examining only intermediates rather than including the transition state for Brz + CzH4 Br- + C2H4Br+comes from an a b initio study12 showing that the transition state is essentially a 3-membered ring plus a Br-. The opposite case was found for Fz, which added in a syn fashion via a nonsymmetrical 4-center transition state.
-
Theoretical Methods Geometries were optimized at the S C F and at the configuration interaction (CI) levels of theory with use of analytic gradient method^.'^*^' The configurations in the CI expansion included single and double substitutions with respect to the Hartree-Fock reference determinant (CISD). The core orbitals were kept frozen (Is on C, Is, 2s, 2p, 3s. 3p on
(7) Vairamani, M.; Srinivas, R.; Mirza, U. A. Org. Mass Spectrom. 1988, 23, 620. (8) Slebocka-Tilk, H.; Ball, R. G.;Brown, R. S. J . Am. Chem. Soc. 1985,
107. 4504. (9) Dewar, M. J. S.; Ford, G. P. J . Am. Chem. Soc. 1979, 101, 783. (10) Cremer, D.; Kraka, E.J . Am. Chem. SOC.1985, 107. 3800. ( I I ) (a) Poirier, R. A.; Merey, P. G.; Yates, K.; Csizmadia, I. J . Mol. Strucr. 1981, 85, 153. (b) Poirier, R. A.; Demare', G. R.; Yates, K.;Csizmadia, I. J. Mol. Strucr. 1983, 98, 137. (12) Yamabe, S.; Minato, T.; Inagaki, S. Chem. Commun. 1988, 532. ( I 3) Pulay, P. In Modern Theoretical Chemisrry; Schaefer, H. F., Ed.; Plenum: New York, 1977; Vol. 4, p 153. (14) (a) Brooks, B. R.; Laidig. W. D.; Saxe, P.; Goddard, J. D.; Yamaguchi. Y.; Schaefer, H. F. J . Chem. Phys. 1980, 72,4652. (b) Rice, J. E.; Amos, R. D.; Handy, N. C.; Lee, T. J.; Schacfer, H. F. J . Chem. Phys. 1986, 85, 963.
J . Am. Chem. Soc.. Vol. 1 1 2, No. 23, I990 8261 Br) as were the 8 virtual orbitals with energies over 20 hartrees. Harmonic vibrational frequencies at the S C F level were obtained analytically as Location of the "transition state' (see Figure 2) from the 2-bromoethyl cation was done by following the eigenvectori6 of the H4-Ci-H3 rocking mode. The smallest basis set employed was of double-{ (DZ) quality, namely the C(9s5p/4s2p) and H(4s/2s) Huzinaga-Dunning basis combined with Dunning's Br(l5slZp5d/8s6p2d) set that was published in an article by Bauschlicher, Schaefer, and Bagus.19 The next basis set consisted of the DZ basis plus the addition of d functions to the heavy atoms (DZ+d), with exponents suitable for describing polarization effects: q ( C ) = 0.75 and a,(Br) = 0.40. The standard designation of this DZ+d basis is C(9~5pld/4sZpld),H(4s/2s), and Br(l5sl2p6d/8~6p3d). The largest basis utilized a triple-{ contracted set with two sets of polarization functions (TZ2P) on the heavy atoms and a D Z set for H with a polarizing set of p functions [ap(H) = 0.751. The exponents of the primitive functions in the T Z and D Z basis sets were the same; the T Z contraction coefficients for C are from ref 18 and those for Br are from ref 20. The pairs of d-function exponents were ad(C) = I A 0 . 3 5 and ad(Br) = 0.54, 0.19. The C(9s5p/5s3p), H(4slp/2slp), and Br( 1 5s 12p7d/ 1Os8p4d) basis set will be referred to as TZ2P for brevity, even though it is only of DZP quality for the hydrogen atoms. The d functions in the S C F calculations were the 5-component spherical harmonic functions, whereas the 6-component Cartesian d functions were used in the CISD calculations due to current program limitations. This should not affect the comparison of the S C F and CISD results in any significant way.
Results and Discussion Figure 2 displays the various stationary points studied. The selected bond lengths and angles are those that are deemed the most reliable, i.e., DZ+d CISD optimized parameters. Table I includes all of the optimized bond distances and angles for the DZ+d SCF, TZ2P SCF, and DZ+d CISD levels of theory. The only noteworthy change in bond angles with respect to level of theory is the increase in the CCBr angle of 1.8O in the 2bromoethyl cation upon inclusion of electron correlation. As a result there is a concomitant closing of the HS-CZ-H6 angle by 1.9'. The primary differences with respect to level of theory are in bond lengths: the increase of about 0.01 A for CISD over S C F for C-H bonds is uniform and quite expected, whereas the C-C and C-Br distances exhibit more variability. Electron correlation decreases the C-Br distance in all cases, indicating a significant contribution to the bonding since correlation usually lengthens bonds. This effect is particularly noticeable in the bromonium ion because electron correlation is generally needed for the description of bridging systems anyway. The improvement of the basis set at the SCF level shortened the C-C bonds by about 0.01 A, again as expected, but the effect of CISD is the shortening of the C-C bond by varying amounts. This is probably due to a hyperconjugative interaction between the empty p orbital on the carbocation center and a "pseudo r" bonding MO of the same symmetry on the methyl group. This same interaction is that which is also used to explain why double bonds are eclipsed with respect to a methyl group hydrogen2' (see Figure 3). The structure of bridged C2H4Br+can be compared with that of 3-membered rings and r-complexes in an attempt to label it. In the bromonium ion, the C-Br distance of 2.025 A is a bit longer than a typical single bond lengthZZof 1.94 A. The C-Br distance from the X-ray determinations is 2.155 A, which is 0.13 A longer than the theoretical structure for the parent. The difference is (15) Saxe, P.; Yamaguchi, Y.; Schaefer, H. F.J . Chem. Phys. 1982, 77, 5647. (16) Baker, J. J . Comput. Chem. 1986, 7, 385. (17) Dunning, T. H. J . Chem. Phys. 1970, 53, 2823. (18) Huzinaga, S. J. Chem. Phys. 1965, 42, 1293. (19) Bauschlicher, C. W.; Schaefer, H.F.; Bagus, P. S. J . Am. Chem. Soc. 1977. 99, 1048. (20) Scuseria, G. E.; Duran. M.; Maclagan. - R. G. A. R.; Schaefer, H.F. J . Am. Chem. SOC.1986, 108, 3248. (21) (a) Radom, L.; Pople, J. A.; Schleyer, P. v. R. J. Am. Chem. Soc. 1972, 94. 5935. (b) Hoffmann, R.; Radom. L.; Pople, J. A.; Schleyer, P. v. R. J. Am. Chem. Sot. 1972. 94. 6221. (22) Huheey, J. E. Inorganic' Chemistry; Harper and Row: New York, 1983; p 545 (1978).
Hamilton and Schaefer
8262 J. Am. Chem. SOC.,Vol. 112, No. 23, 1990 2-bromoethyl cation
bromonium ion
/'
H
111.8'
trans I -bromoethyl cation
cis I-bromoeihyl cation
[ransition state
Br \
B;
Figure 2. The isomers of C2H4Brt considered in this study with DZ+d CISD optimized geometry parameters of note. The values for the transition-state structure are from the DZ+d SCF optimized geometry.
Figure 3. The hyperconjugative interaction between the empty p orbital on the cation center with the occupied methyl group MO of A symmetry.
due to the fact that the adamantyl derivative is expected to allow less distortion of the ethylenic plane, which results in the intermediate resembling a *-complex more, with an attendant weaker bonding and larger C-Br distance. The bromonium C-C distance of 1.455 A is between the usual values of 1.34 A for C-C and 1.54 A for C-C. The adamantyl substituted C-C bond length of 1.497 A is rather extreme for a *-complex (the C-C bond in cyclopropane is 1.497 at the 6-31G* SCF level!I0). Our first thought was that this long distance could be explained by steric repulsion, but the unbrominated adamantylideneadamantane C-C distance is 1.336 A.23 This is an increase of 0.161 A upon bromination, which is more indicative of a m m p l e x . The X-ray structure thus sheds no light on the nature of the ethylenebromonium ion. The ab initio C-C length for C2H4Br+upon bromination is 1.455 A, which is slightly above the up er bound for *-complexes given by lttel and Ibers (1 -36-1.44 ).24 The bending of the CH, groups away from the former C2H4plane by about 17.4' is intermediate between ethylene (OO) and cyclopropane (30'). The typical distortion of C2H4 for planarity for *-complexes is 16-20O." Even the strained adamantyl compound shows an out-of-plane bend of 16.4'. The last indicator of u or T character to be considered is charge transfer. The bonding in *-complexes is primarily from the ethylene K orbital to an empty adduct orbital. The stronger a-complexes form with back-donation of electron density from the adduct orbital of proper symmetry to the A* MO, with little net charge transfer. From a charge analysis derived from the dipole moment derivative^,^^ the positive charge does not reside on the Br atom at all, but on the C atoms. This is also qualitatively the same charge distribution that comes from a Mulliken population analysis. This means that very little back-donation of electron density has occurred, if any. If forced to choose, it seems
K
(23) Swen-Walstra, S. C.; Visser, G. J . J . Chem. Soe. D 1971, 82. (24) Ittel, S. D.; Ibers, J . A. Ado. Organomet. Chem. 1976, 14, 33. (25) Cioslowski. J.; Hamilton, T. P.; Scuseria, G.; Hess, 8. A.; Hu, J.; Schaad, L. J.: Dupuis, M. J . Am. Chem. Soc. 1990. 112, 4183.
that the bridged C2H4Br+is a strong *-complex rather than a a-complex. At first glance the 2-bromoethyl cation (which will turn out to be a transition state) has a little more multiple bond character than the bromonium ion, but it must be realized that the bond arises from the overlap of sp2 and sp3 orbitals. The C-Br bond is significantly shortened to the point where it is even a little shorter than the typical C-Br bond. There is a great deal of repulsion between the eclipsed C-H and C-Br bonds: the CCBr angle is 9.7O larger than the sp3 value of 109.5O! This results in more p character in the bonds to the CH2 group and the HCH angle decreases 8.5O from the tetrahedral ideal. The bending of the CH2group is enhanced by the interaction of the pseudo-n orbitals with the empty p orbital on the cation center. The resulting loss of electron density in the C-H bonding M O gives rise to an increase in the C-H bond distance of about 0.015 A. The large amount of s character remaining in the bond to the Br atom explains why it is shortened so much. The I-bromoethyl cations are very similar to each other. Both have slightly increased C-C bond lengths when compared to all of the other structures. This is due to the C-Br acquiring a fair amount of double bond character. As a matter of fact, the C-C bond may be considered as a single bond in which the short distance arises from overlap of sp2 and sp3. There is naturally more repulsion in the form where C-Br and C-H are eclipsed (resulting in larger bond angles) than in the case where C-H and C-H are eclipsing. The pseudo-* orbital of the methyl group is better able to interact with the carbocation center when the partially occupied C-Br r orbital is trans to it; however, this cannot overcome the repulsion to make the trans isomer the lower energy one (see Table 11). The partial ?r bond from the Br to the cation center makes the interaction between the pseudo-* methyl M O and the cation center weaker for both the cis and trans cases. This is reflected in the fact that C-H bonds are shorter and the angles are more reminiscent of sp3, especially for the trans isomer. We see a definite preference for the partial double bond to eclipse the methyl by 1.5 kcal/mol, an effect that has long been noted.26 Finally, the transition state resembles bromoethene with a proton sitting on the *-electron cloud which is polarized toward the bromine-substituted end. The C-C distance exhibits more double bond character by virtue of its length than those in any of the other structures. (26) Hehre. W. J.: Salem, L. Chem. Commun. 1973, 743.
J . Am. Chem. SOC.,Vol. 112, No. 23, 1990 8263
Structure and Energetics of C2H4Br+ Table I. Optimized Geometrical Parameters for C2H4Br+Stationary Points
DZ+d S C F 1.442 1.074 2.053 107.7 117.9
DZ+d CISD 1.455 1.084 2.025 108.7 117.8
TZ2P S C F 1.434 1.073 2.060 107.3 118.2
2-bromoethyl cation
1.447 1.08I 1.082 I .094 1.909 123.2 119.1 106.2 103.9 116.4
1.434 1.091 1.092 1.109 1.902 123.2 119.0 105.5 101.0 118.2
1.434 1.08 1 1.082 1.096 1.907 123.1 119.0 106.3 103.6 116.6
I-bromoethyl cation (cis)
I .466 1.091 1.078 1.079 1.776 125.3 119.1 1 1 3.9 107.0 105.7 1.473 1.079 1.089 1.079 1.779 123.9 121.0 1 1 1.3 108.7 106.4 1.405 1.664 1.156 1.080 1.079 1.078 1.892 119.7 121.2 106.6 110.6 117.4 121.0 105.7 109.5
1.463 1.101 1.087 1.090 1.770 125.1 119.4 1 14.2 107.1 104.8 1.470 1.088 1.098 I .090 1.772 123.9 121.1 111.3 109.1 105.7
1.454 1.093 1.077 1.079 1.779 125.1 119.6 114.2 106.8 105.1 1.462 1.079 1.090 1.079 1.782 123.7 121.4 111.6 108.4 106.1
bromonium ion
.yC H
B;
I-bromoethyl cation (trans) Br
H
transition state
Table 11. Total and Relative Energies (kcal/mol) of Various C,HABrt Isomers
method DZ S C F DZ+d S C F TZ2P S C F DZ+d C l S D DZ+d C l S D + Davidson 'Energics at
bromonium ion 2-bromoethyl cation -2650.02486 -2650.001 37 0.0 14.7 -2650. I 1376 -2650.07890 0.0 21.9 -2650. I3325 -2650.09857 0.0 21.8 -2650.591 90 -2650.54740 0.0 27.9 -2650.64660 -2650.59980 corr 0.0 29.4 the DZ+d S C F optimized geometry.
structure I-bromoethyl cation (cis) -2650.02444 0.3 -2650. I 1037 2.1 -2650. I 2902 2.7 -2650.58746 2.8 -2650.64227 2.7
I-bromoethyl cation (trans) -2650.02264 1.4 -2650.10822 3.5 -2650.1 2666 4.1 -2650.58505 4.3 -2650.63982 4.3
transition state -2649.991 34 21.0 -2650.07910 21.8 -2650.09988" 20.9' -2650.551 37' 25.4" -2650.60438" 26.5'
8264 J . Am. Chem. Soc., Vol. 112. No. 23, 1990
Hamilton and Schaefer
Table 111. DZ+d SCF Vibrational Frequencies (cm-I) and IR Intensities (km/mol) for C2H4Br+Ssystems 1-bromoethyl 2-bromoethyl I-bromoethyl bromonium ion cation cation (cis) cation (trans)
transition state
3499 3489 3368 3365 I663 1596 1318 I296 I222 1204 1024 95 1 896 45 5 413
45
3454 3329 3190 3153 1632 I446 I424 1417 I141 1113 I047 649 581 334 264i
0
I 18 IO II 0 6 52 1
0 I 2
75 40
31 6 32 96 0 77 15 4 66 4 11
23 IO 13 12
3403 3378 3231 3170 1609 I569 1483 1455 1243 1110 I058 831 743 359 I09
Table I I records the energies of the isomers at the level of theory in which thcy were optimized. The only exception was the one labeled as “transition state” in Figure 2, because the optimization in C, symmetry past the DZ+d SCF level proved insurmountable. The cvaluation of energies at higher levels of theory with DZ+d S C F geomctries performed on the other isomers showed that optimization lowered the DZ+d CISD energy by only 0.2-0.4 kcal/mol for the ethyl cations and 4.7 kcal/mol for the bromonium ion. The DZ+d CISD optimized energy will most likely be lower than the singlc point energy by an amount in this range. Also, the effect of completely optimizing all of the geometry parameters had only a small effect on the DZ S C F results of ref I I . The S C F vibrational frequencies and IR intensities for the differcnt stationary points are given in Table 111 and are mainly used to characterize the potential energy surface, since there are no cxpcrimcntal vibrational spectra. For the adamantyl-substituted compound,@the lowest two vibrational frequencies of 421 and 403 cm-I may correspond to the C-Br stretch and bend parallel to the C-C bond, as the C2H4Br+harmonic frequencies are 455 and 413 cm-I with respectable 1R intensities. After taking into account that S C F frequencies are typically 10% too high, the C-C stretching frequency of 1500 cm-’ is a bit below the range of 1640-1 680 cm-’ observed for double bonds, which is consistent with a n-complex picture. All energies are reported relative to the cyclic ion which is a global minimum throughout this study. The 2-bromoethyl cation is a transition state that is nearly 30 kcal/mol above the bromonium ion, which by coincidence is the prediction from STO-3G S C F theory.” The imaginary vibrational mode of the 2bromoethyl cation is a torsion; however, all attempts to follow it led to a spontaneous collapse of the C-Br bond toward the middle to give thc cyclic ion. This unusual feature has been noted in previous theoretical work on other C2H,X+ systems*’ and results in the assertion that the 2-bromoethyl cation will not be observed in the laboratory. This removes the suspicion that it could be a very shallow minimum,” the doubt being due to incomplete geometry optimization. The cis-I-bromoethyl cation is indeed a minimum at all levels of theory and is very nearly equal in energy to the bromonium ion. Ab initio agrees amazingly well with experiment on these two dissimilar species. Once the difference in zero-point vibrational energy is taken into account, lowering the cis-I-bromoethyl cation by 1.3 kcal/mol relative to the bromonium ion, our best value for the energy difference is 1.5 kcal/mol. The energy difference is remarkably insensitive to basis set or electron correlation, leading to the conclusion that the value of 1.4 kcal/mol given by Berman et al. is accurate.8 While the Davidson correction negligibly affects the energy differences for the two minima, the actual geometry dependent effects of quadruple excitations may be significant, given the rather small energy difference between the bridged and cis ~~~~
12 6 9 52 13 23 12 101 17 0
I52 5 42 1 I
340 1 3362 3259 3189 1595 1574 1516 1452 1 I93 1 I67 1105
754 718 356 209i
14 3 II 41 2 21 63 39 1
35 60 76 11
6 15
3463 3363 3336 2710 1648 1483 1439 1360 I264 1101
985 680 619 355 392i
30 16 13 123 IO 32 23 IO 18 2 16 1
16 25 209
isomers. The trans-I-halo isomer is a transition state with an imaginary mode that twists in the direction of the cis isomer and is only slightly higher in energy. Presumably, this indicates that rotation about the C-C bond is nearly unhindered, giving credence to the assertion that the C - C bond is merely a short single bond. Beauchamp’s group8 reported that the cyclic and acyclic minima did not interconvert under their experimental conditions (room temperature), and indeed the structure that is labeled “transition state” presents an energy barrier of 25 kcal/mol for H migration. The vibrational frequency analysis and structure make it clear that this is the previously unreported transition state for the interconversion. The other conceivable pathway that goes through opening the bromonium ion to the 2-bromoethyl ion, followed by H migration via a stationary point with at least 2 imaginary modes, is not feasible. The final topic of discussion will be on the effects expected when methyl groups are substituted for hydrogen. There has been a great deal of experimental study on this question, and some low level ab initio work on methyl-substituted halonium ions with halogens smaller than b r ~ m i n e . ~ ’ * * ~ The replacement of an H by CH3 will induce an asymmetry in the bromonium ion. The methyl will stabilize the positive charge on the (Y carbon and the Br atom will be shifted toward the unsubstituted carbon atom. This also agrees with Dewar and Ford’s9 argument that the r orbital will be polarized toward the primary carbon, so the Br atom will be shifted that direction in the r-complex. The open ions will all be stabilized by roughly the same amount by a single methyl substitution, whereas the bridged isomer will not benefit to the same degree. Therefore, the 2-bromoisopropyl cation may become the global minimum. The best chance for observing a methylated derivative of the open 2-bromoethyl cation is to place both methyl groups on the (Y carbon. This should favor concentrating the charge on the a carbon enough so that the open form may be more stable. Given the 30 kcal/mol difference in the parent ion energies this is not likely. The 1,2-dimethyl isomers that arise from bromination of cis- and trans-2-butenes should be symmetric bridged intermediates. The energy of the open ions produced by ring opening should be decreased relative to the bridged, but not as much as for the 1,l-dimethyl ion. In superacid media, the 1,2-dimethyl isomers undergo H and CH3 shifts to give the bridged 1,l-dimethylethylenebromonium ion. When a third methyl is added, the qualitative trends seen for the 1,l-dimethyl case should be expected, with the Br atom experiencing longer excursions toward the (Y carbon. The intermediate from the bromination of 2,3-dimethyl-2-butene should be symmetric, with the movement of the Br from the (Y carbon to the (3 carbon (and vice versa) being more facile. The experimental observations in superacid media5 support the proposition that the intermediates are all bridged for the bromine case, whereas
~
(27) Hehre. W. J.; Hiberty, P. C. J . Am. Chem. SOC.1974, 96, 2665.
(28) Yamabe, S.; Tsuji, T.; Hirao, K . Chem. Phys. Leu. 1988, 146. 236.
J . Am. Chem. SOC.1990, 112, 8265-8268
8265
sitionstate structure in which a H atom is bridging, with a barrier of 25 kcal/mol. The best description of the bridged ion is as a strong r-complex. It must be reiterated that alkyl substitution can have a great effect on the energetic^,^'.^^ so the results from this study must be applied to other bromonium systems with great care. The ab initio study of halonium ions in larger alkenes would be highly desirable because of the possible role of these structures in the charge conduction mechanism of doped p~lyacetylene.~~
open 2-halo cations may be possible for chlorine and fluorine.
Conclusions The two lowest minima of the five structures examined in this study are nearly degenerate, with the bromonium ion coming out lower in energy at all levels of theory. Given the insensitivity of this relative energy to basis set and correlation, we expect that these results, which are in excellent agreement with experiment, will not change substantially upon improving the theoretical treatment. The prediction that there is no bound 2-bromoethyl cation is confirmed conclusively. The rotation about the C-C bond in the I-bromoethyl cation is nearly free. Since the I-bromo form is 25 kcal/mol lower in energy than the 2-bromo isomer, protonation of bromoethene will occur only in a Markovnikov fashion. The interconversion of the two minima occurs through a tran-
Acknowledgment. This research was supported by the U. S. Department of Energy, Office of Basic Energy Sciences, Division of Chemical Sciences, Fundamental Interactions Branch, Grant DE-FG09-87ER138 1 1 . (29) Borman,
S. Chem. Eng. News 1990. 68 (19), 53.
77SeNMR and Crystallographic Studies of Selenazofurin and Its 5-Amino Derivative Barry M. Goldstein,*.+ Scott D. Kennedy,+ and William J. Hennen”8 Contribution from the Department of Biophysics, University of Rochester Medical Center, Rochester, New York 14642, and Department of Chemistry, Brigham Young University, Provo, Utah 84602. Received September 1I I989 ~
Abstract: Studies presented here examine the hypothesis that the close Se-01’ contact observed in the chemotherapeutic agent selenazofurin is electrostatic in origin. The crystal structure and 77Sespectrum of selenazofurin are compared with those of the 5-amino derivative. The 77Sespectrum of selenazofurin shows a doublet of doublets at 774.2 ( 5 ) ppm downfield of dimethyl selenide. A two-bond J coupling of 44.3 Hz is observed between Se and the selenazole heterocycle proton H5. A three-bond J coupling of 3.2 Hz is observed between Se and the furanose proton HI’. These observations are consistent with a positively charged selenium in a partially delocalized selenazole ring, in agreement with earlier X-ray studies. The 77Sesignal for the 5-amino derivative appears as a doublet at 687.5 (5) ppm relative to dimethyl selenide and 87 ppm upfield from that observed for the parent compound. A two-bond J coupling between Se and HI’ of 4.2 Hz is observed in the 5-amino derivative. The crystal structure of 5-aminoselenazofurin shows an unusually high glycosidic torsion angle with an increase in the S e a l ’ distance relative to that found in the parent compound. These findings are consistent with a decrease in positive charge on the selenium in the 5-amino derivative, resulting in a decrease in the attractive component of the Se-01‘ interaction and a shift in conformation about the C-glycosidic bond.
Introduction Selenazofurin (Figure 1a, 2-~-~-ribofuranosylselenazole-4carboxamide, NSC 340847) is a C-glycosyl selenazole nucleoside demonstrating a wide variety of antitumor’ and antiviral2 activities. The crystal structures of both selenazofurin (Figure 1 b) and its a anomer show an unusual conformational f e a t ~ r e . ~In each structure, the distance between the selenium atom and the furanose oxygen 01’is significantly less than the sum of the Se and 0 van der Waals radii. Similar close intramolecular selenium-oxygen contacts have been noted in crystal structures of two selenophene derivative^,^.^ and a number of close intermolecular seleniumnucleophile contacts have also been catalogued.6 Selenazofurin is the selenium analogue of the antitumor agent tiazofurin. In tiazofurin, the selenium atom of selenazofurin is replaced by a sulfur, forming a C-glycosyl thiazole nucleoside. Crystal structures of tiazofurin and seven analogues also show close heteroatom-oxygen Preliminary computational studies in the thiazole nucleosides suggest that these close S - 0 contacts result from an electrostatic interaction between a positively charged sulfur and negatively charged furanose oxygen.I0 Author to whom correspondence should be addressed. ‘University of Rochester Medical Center. ‘Current address: Parish Chemical Company, 145 North Geneva Road, Oram, UT 84057. t Brigham Young University.
0002-7863/90/ 15 12-8265$02.50/0
However, the electronic structure of the selenazole heterocycle has not been widely studied. Crystallographic data on true sel(1) (a) Parandoosh, Z.; Robins, R. K.;&lei, M.; Rubalcava, B. Biochem. Biophys. Res. Commun. 1989, 164, 869-874. (b) Sokoloski. J. A.; Blair, 0. L.; Sartorelli, A. C. Cancer Res. 1986. 46, 2314-2319. (c) Boritzki, T. J.; Berry, D. A.; Besserer, J. A.; Cook, P. D.; Fry, D. W.; Leopold, W. R.; Jackson, R. C. Biochem. Pharmacol. 1985.34, 1109-1 114. (d) Bergtr, N. A.; Berger, S.J.; Catino, D. M.; Petzold, S.J.; Robins, R. K. J. Clin. Inuesr. 1985, 75, 702-709. (e) Jayaram, H. N.; Ahluwalia, G. S.;Dion, R. L.; Gebeyehu, G.;Marquez, V. E.; Kelley, J. A.; Robins, R. K.; Cooney, D.A,; Johns, D. G. Biochem. Pharmacol. 1983,32,2633-2636. ( f ) Srivastava, P. C.; Robins, R. K. J . Med. Chem. 1983, 26, 445-448. (g) Streeter, D. G.; Robins, R. K. Biochem. Biophys. Res. Commun. 1983, 115, 544-550. (2) (a) Sidwell, R. W.; Huffman, J. H.; Call, E. W.; Alaghamandan, H.; Cook, P. D.; Robins, R. K. Anlimicrob. Agents Chemorher. 1985, 28, 375-377. (b) Kirsi, J. J.; North, J. A.; McKernan, P. A.; Murray, B. K.; Canonico, P.G.; Huggins, J. W.; Srivastava, P. C.; Robins, R. K. Antimicrob. Agents Chemother. 1983, 24, 353-361. (3) Goldstein, B. M.; Takusagawa, F.; Berman, H. M.; Srivastava, P. C.; Robins, R. K. J. Am. Chem. SOC.1985, 107, 1394-1400. (4) Nardelli, M.; Fava, G.; Giraldi, G. Acra Crystallogr. 1962, IS, 737-746.
( 5 ) (a) Susse, P.; Kunz, G.;Luttke, W. Narunvissenschafren 1973, 60. 49-50. (b) von E l k , H. Bull. Soc. Chim. Fr. 1955, 106, 1429-1433. (6) Ramasubbu, N.; Parthasarathy, R. Phosphorus Sul/ur 1987, 31, 22 1-229. (7) Goldstein, B. M.; Takusagawa, F.; Berman, H. M.; Srivastava, P.C.; Robins, R. K. J. Am. Chem. Soc. 1983, 105, 7416-7422. (8) Goldstein, B. M.; Mao, D. T.; Marquez, V. E . J . Med. Chem. 1988, 31, 1026-1031.
0 1990 American Chemical Society
